pubs.acs.org/Langmuir © 2009 American Chemical Society
Atomic Layer Deposition and Abrupt Wetting Transitions on Nonwoven Polypropylene and Woven Cotton Fabrics G. Kevin Hyde,†,§ Giovanna Scarel,† Joseph C. Spagnola,‡ Qing Peng,† Kyoungmi Lee,† Bo Gong,† Kim G. Roberts,†,§ Kelly M. Roth,† Christopher A. Hanson,† Christina K. Devine,† S. Michael Stewart,† Daisuke Hojo,† Jeong-Seok Na,† Jesse S. Jur,† and Gregory N. Parsons*,†,‡ † Department of Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, North Carolina 27695, and ‡Department of Materials Science and Engineering, North Carolina State University, Raleigh, North Carolina 27695. §Alditri Technologies Inc., Raleigh, NC 27617
Received July 31, 2009. Revised Manuscript Received September 7, 2009 Atomic layer deposition (ALD) of aluminum oxide on nonwoven polypropylene and woven cotton fabric materials can be used to transform and control fiber surface wetting properties. Infrared analysis shows that ALD can produce a uniform coating throughout the nonwoven polypropylene fiber matrix, and the amount of coating can be controlled by the number of ALD cycles. Upon coating by ALD aluminum oxide, nonwetting hydrophobic polypropylene fibers transition to either a metastable hydrophobic or a fully wetting hydrophilic state, consistent with well-known Cassie-Baxter and Wenzel models of surface wetting of roughened surfaces. The observed nonwetting/wetting transition depends on ALD process variables such as the number of ALD coating cycles and deposition temperature. Cotton fabrics coated with ALD aluminum oxide at moderate temperatures were also observed to transition from a natural wetting state to a metastable hydrophobic state and back to wetting depending on the number of ALD cycles. The transitions on cotton appear to be less sensitive to deposition temperature. The results provide insight into the effect of ALD film growth mechanisms on hydrophobic and hydrophilic polymers and fibrous structures. The ability to adjust and control surface energy, surface reactivity, and wettability of polymer and natural fiber systems using atomic layer deposition may enable a wide range of new applications for functional fiber-based systems.
I. Introduction Natural and synthetic fibrous materials are finding a wide range of new applications including advanced filtration,1 composite reinforcement,2 as membranes for protection and biological species capture,3 tissue engineering and biological adhesive scaffolds,4,5 and substrates for printed and flexible electronic circuits and sensors.6,7 Fibrous structures are being adopted for an expanding range of applications such as catalytic and photocatalytic mantles,8-10 materials for biological product and byproduct separations, protective and biocidal garments, and flexible and low-cost energy harvesting devices. For most applications of fibrous materials, the physical structure and chemical composition of the fiber surface dictate the device functionality and performance. New techniques to uniformly adjust and modify surface properties, while maintaining *To whom correspondence should be addressed. (1) Wente, V. Ind. Eng. Chem. 1956, 48, 1342–1346. (2) Bledzki, A. K.; Gassan, J. Prog. Polym. Sci. 1999, 24, 221–274. (3) Gibson, P.; Schreuder-Gibson, H.; Rivin, D. Colloids Surf., A 2001, 187, 469–481. (4) Bhattarai, S. R.; Bhattarai, N.; Yi, H. K.; Hwang, P. H.; Cha, D. I.; Kim, H. Y. Biomaterials 2004, 25, 2595–2606. (5) Hyde, G. K.; McCullen, S. D.; Jeon, S.; Stewart, S. M.; Jeon, H.; Loboa, E. G.; Parsons, G. N. Biomed. Mater. 2009, 4 (2), 025001, 1-10. (6) Karaguzel, B.; Merritt, C. R.; Kang, T.; Wilson, J. M.; Nagle, H. T.; Grant, E.; Pourdeyhimi, B. J. Text. Inst. 2009, 100, 1–9. (7) Shim, B. S.; Chen, W.; Doty, C.; Xu, C.; Kotov, N. A. Nano Lett. 2008, 8, 4151–4157. (8) Kemell, M.; Pore, V.; Ritala, M.; Leskel€a, M.; Linden, M. J. Am. Chem. Soc. 2005, 127, 14178–14179. (9) Santala, E.; Kemell, M.; Leskel€a, M.; Ritala, M. Nanotechnology 2009, 20, 1–5. (10) Staira, P. C.; Marshall, C.; Xiong, G.; Feng, H.; Pellin, M. J.; Elam, J. W.; Curtiss, L.; Iton, L.; Kung, H.; Kung, M.; Wang, H.-H. Top. Catal. 2006, 39, 181–186.
2550 DOI: 10.1021/la902830d
the desirable properties of the original fiber matrix, will be important for many new device and system applications. Methods to modify surfaces to achieve superhydrophobic or superoleophobic behavior have been of interest for some time,11-14 in particular in fiber- and fabric-based systems.15-17 The dynamic wetting properties of a surface, which are associated with advancing and receding contact angles,11 result from a combination of surface chemistry and surface roughness and texture.18,19 A vast number of wet chemical techniques have been developed to control surface properties of fibers and fabrics,20 and new approaches continue to emerge.21-24 Vapor phase atmospheric pressure plasmas, in either inert gas or with chemical additives,25,26 (11) Chen, W.; Fedeev, A. W.; Hsieh, M. C.; Oner, D.; Youngblood, J.; McCarthy, T. J. Langmuir 1999, 15, 3395–3399. (12) Dorrer, C.; Ruhe, J. Soft Matter 2009, 5, 51–61. (13) Ma, M.; Mao, Y.; Gupta, M.; Gleason, K. K.; Rutledge, G. C. Macromolecules 2005, 38, 9742–9748. (14) Tuteja, A.; Choi, W.; Ma, M.; Mabry, J. M.; Mazzella, S. A.; Rutledge, G. C.; McKinley, G. H.; Cohen, R. E. Science 2007, 318, 1618–1622. (15) Bhat, K.; Heikenfeld, J.; Agarwal, M.; Lvov, Y.; Varahramyan, K. Appl. Phys. Lett. 2007, 91, 1–3. (16) Choi, W.; Tuteja, A.; Chhatre, S.; Mabry, J. M.; Cohen, R. E.; McKinley, G. H. Adv. Mater. 2009, 21, 1–6. (17) Wu, H.; Zhang, R.; Sun, Y.; Lin, D.; Sun, Z.; Pan, W.; Downs, P. Soft Matter 2008, 4, 2429–2433. (18) Cassie, A. B. D.; Baxter, S. Trans. Faraday Soc. 1944, 40, 546–551. (19) Wenzel, R. N. Ind. Eng. Chem. 1936, 28, 988–994. (20) Bajaj, P. J. Appl. Polym. Sci. 2002, 83, 631–659. (21) Hyde, K.; Dong, H.; Hinestroza, J. P. Cellulose 2007, 14, 615–623. (22) Hyde, K.; Hinestroza, J.; Rusa, M. Nanotechnology 2005, 16, S422–S428. (23) Mahltig, B.; Haufe, H.; Bottcher, H. J. Mater. Chem. 2005, 15, 4385–4398. (24) Muller, K.; Quinn, J. F.; Johnston, A. P. R.; Becker, M.; Greiner, A.; Caruso, F. Chem. Mater. 2006, 18, 2397–2403. (25) McCord, M. G.; Hwang, Y. J.; Hauser, P. J.; Qiu, Y.; Cuomo, J. J.; Hankins, O. E.; Bourham, M. A.; Canup, L. K. Text. Res. J. 2002, 72, 491–498. (26) Tsai, P. P.; Roth, L. C. W. J. R. Text. Res. J. 1997, 67, 359–369. (27) Kemell, M.; Ritala, M.; Leskel€a, M.; Groenen, R.; Lindfors, S. Chem. Vap. Deposition 2008, 14, 347–352.
Published on Web 10/02/2009
Langmuir 2010, 26(4), 2550–2558
Hyde et al.
Article
as well as chemical vapor deposition13 and atomic layer deposition8,27-30 are of interest. Atomic layer deposition is particularly attractive because it uses a binary set of self-limiting reactions that can produce highly uniform coatings on complex surface geometries, where the thickness is controlled at the (sub)monolayer level by the number of ALD reaction cycles. ALD reaction chemistry on polymers differs from that on dense inorganic solids, and the detailed differences depend on the nature of the polymer and the deposition reactants.31,32 While changes in surface roughness and surface wetting have been observed after ALD coating of polymer films,31,32 surface wetting characteristics of ALD-coated fibers and textile media have not previously been studied in detail. The surface wetting is of interest, in part, because the wettability of a fabric material has a strong influence on its performance, for example, in liquid and aerosol filtration. Wetting properties of fibers are modulated by surface chemical species. The ability to control species can allow further chemical functionalization that can significantly expand the range of potential applications. In this work, we analyze the uniformity of low-temperature (8%. Decomposition of cotton occurs at temperatures greater than 150 C. Melt-blown polypropylene fiber mats were prepared at NC State University, College of Textiles, and used as received. Polypropylene thin films were cast on silicon substrates to approximate a smooth textured surface. The polypropylene solution was prepared by dissolving polypropylene beads in decahydronaphthalene to create a 1 wt % solution. The solution was heated to 165 C for at least 1 h to complete the dissolution of the polypropylene. An 80 μL amount of the solution was placed onto a silicon substrate heated to 100 C to evaporate the solvent. The root-mean-square roughness of the resulting film measured over 2 μm2 μm area is 6-8 nm. Woven cotton fabrics, with a 31 twill structure, were obtained from Textile Innovators and used as received (bleached and mercerized). To approximate the chemical surface of the cotton fibers in a planar form, cellulose films were prepared by spincasting a solution containing microcrystalline Avicel cellulose on silicon substrates. The solution preparation and spin-cast process followed a procedure reported elsewhere.33 Aluminum oxide atomic layer deposition was performed in one of three homemade hot-wall stainless viscous flow tubular reactors, all of similar design.29,30,34 Each reactor used is comprised of a stainless steel tube ∼75 cm long and ∼3.5 cm in diameter. Trimethylaluminum (TMA, Al(CH3)3) and deionized water (H2O) were used as Al precursor and oxidant, respectively. The ALD process proceeds through a series of binary self-limiting A/B reaction cycles, where, for example, TMA reacts with a hydroxylterminated surface to form aluminum-methyl species bound to either one or two surface oxygen atoms (reaction A) followed by water reacting with the Al-methyl groups to form a (sub)monolayer of Al2O3 with hydroxyl surface termination (reaction B). A continuous flow of ultrahigh-purity argon was used and acted as precursor carrier and purge gas. Each reactant was delivered through a dedicated 1/4 in. stainless steel tube. The deposition tube and gas lines were independently temperature controlled using heating jackets. The maximum service temperature of polypropylene is typically 125 C and ∼150 C for cotton. Therefore, for our studies, the ALD process temperature is limited to a maximum of 120 C. The TMA was obtained from STREM Chemicals and used as received. The total gas flow rate was typically ∼200 standard cubic centimeters per minute (sccm), and the operating pressure was maintained at 1-2 Torr using a rotary mechanical pump. Pressure was monitored by a Baratron pressure gauge (MKS Instruments Inc.). The pressure increase during each reactant gas pulse was between 50 and 100 mTorr. X-ray diffraction analysis confirms that the alumina is amorphous in the as-deposited state. Advancing and receding water contact angle on the various fabrics was measured using a Model 200 Rame-Hart contact angle goniometer. All reported contact angle values are obtained within 1 h or less after deposition in order to reduce possible effects of adventitious carbon on the surface. After proper tool alignment, the water droplet was allowed to stabilize for about 1 min on the fiber before recording the measurement. The reported values correspond to the advancing contact angle. Because of the rough surface on the fibers, the contact angle was evaluated by the droplet shape close to the droplet/fiber contact (see for example Figure 7a). The reported values correspond to averages from (33) Song, J.; Liang, J.; Liu, X.; Krause, W. E.; Hinestroza, J. P.; Rojas, O. J. Thin Solid Films 2009, 517, 4348–4354. (34) Hyde, G. K.; Park, K. J.; Stewart, S. M.; Hinestroza, J. P.; Parsons, G. N. Langmuir 2007, 23(19), 9844–9849.
DOI: 10.1021/la902830d
2551
Article
Hyde et al.
Figure 1. SEM images melt-blown polypropylene fiber mats showing smooth surface of the initial fibers. Typical fiber diameters range between 0.5 and 10 μm.
Figure 2. (a) Example of the “peeling” process to achieve thin polypropylene fiber samples from a thicker fibrous mat. Peeled films were used for infrared absorbance spectra analysis. (b) Spectra obtained from an uncoated “peeled” polypropylene fiber matrix (bottom) compared to the same matrix after 200 A/B cycles of TMA/H2O to produce Al2O3, deposited at 90 C (middle). The difference spectrum is also provided (top). between 3 and 10 measurements of each sample, with typical uncertainty (one standard deviation) on the nonwetting surfaces of less than (10. Characterization of the fiber surfaces was conducted using a FEI Phenom scanning electron microscope (SEM) which analyzes backscattered electrons and operates with a 5 kV accelerating voltage. A Hitachi S-3200 SEM was used for higher resolution imaging. To reduce charging effects, the fiber and film samples with and without ALD processing were sputtercoated with 120. After coating with a small number of A/B cycles, the films remain nonwetting. Water droplets on the nonwetting samples remained stable and were visible until the fluid evaporated. As the number of A/B cycles is increased, samples prepared in either orientation undergo an abrupt transition to a wetting state, where they readily soak up water. The nonwetting/wetting transition occurs at approximately the same number of A/B cycles for both sample orientations. It is important to note that the wetting and nonwetting nature of the ALD-coated polypropylene fibers is found to be nearly unchanged after exposure to lab air for more than 120 days. To better understand the observed wetting transition on the polypropylene fibers, further experiments were performed at various deposition temperatures in another reactor system, and results for samples coated at 60 and 90 C are shown in Figure 6a using a “flow-over” sample orientation. Deposition was performed using TMA/Ar/H2O/Ar=1/30/1/60 s, which is an overall longer cycle time than that shown in Figure 5 but still results in ALD growth. Here, high-temperature deposition conditions showed no evidence of a nonwetting/wetting transition. The nonwetting characteristic was observed for temperatures up to 120 C even after as many as 1000 A/B cycles (data points not shown). At 60 C, an abrupt nonwetting/wetting transition occurs at ∼50 A/B cycles. This result shows that the nonwetting/wetting transition depends strongly on deposition temperature. This effect is consistent between Figures 5 and 6 albeit the transition is observed to occur at different number of cycles, which is most likely a result of some variation in detailed reactor condition. A control experiment to confirm that the wetting properties of the fiber systems are related to film deposition was completed by performing a “double deposition”. In this experiment, hydrophilic samples formed with 100 A/B cycles at 60 C were recoated at 90 C, and hydrophobic samples originally coated with 100 A/B cycles at 90 C were coated with another layer at 60 C. Surprisingly, we find the second layer generally does not affect the original surface wetting conditions. When a hydrophilic sample coating at 60 C with 100 cycles is coated with an additional 100 A/B cycles at 90 C, the surface maintains its original wetting state. Similarly when a nonwetting sample formed with 100 A/B cycles at 90 C is coated at 60 C, the surface remains nonwetting. Exceptions do occur when the original nonwetting samples are formed close to the nonwetting/wetting transition point. From this, it can be concluded that the transition observed is most likely a result of nucleation effects of the Al2O3 on the polymer fiber. Langmuir 2010, 26(4), 2550–2558
Hyde et al.
Article
Figure 8. A general wetting diagram of cos(θ*) vs cos(θ) obtained from contact angle on fibers and films for polypropylene coated with various number of A/B cycles of Al2O3. Green points correspond to deposition of aluminum oxide at 90 C; red points: 75 C; and blue points: 60 C.
Figure 7. (a) Image of a water droplet on a nonwetting cotton surface after two A/B cycles of aluminum oxide. (b) Transition of water contact angle on cotton fibers and cellulose films as a function of ALD cycles at various deposition temperatures. One cycle consisted of TMA/Ar/H2O/Ar: 1/30/1/60 s.
As the fiber material surface changes from naturally hydrophobic to wetting, the fibrous mat will transition to a condition that appears hydrophobic but is more accurately described as a metastable Cassie-Baxter state. The metastable state tends to result in droplet adhesion to the surface, with a corresponding increase in contact angle hysteresis (advancing - receding). To examine this transition, water adhesion was tested on the coated polypropylene fibers by gently placing a droplet on the fiber surface and tilting the samples vertically. Water droplets readily rolled off of uncoated polypropylene fibers as well as fibers coated with a small number of ALD cycles. Increasing the number of cycles at 90 C resulted in the droplets adhering to the surface. For the polypropylene coated at 60 C, increasing the number of cycles resulted in droplet adhesion, just before the transition to full wetting. In Figure 6, for example, nonwetting samples formed at 60 C near the nonwetting/wetting transition point showed water adhesion upon tilting, and some showed slow water absorption (within a few minutes), also indicative of the metastable state. Also, fiber materials that remain nonwetting after high temperature ALD processing also appear in the metastable Cassie-Baxter state. The detailed growth mechanisms on the polymer surface play a crucial role in defining the surface chemical composition and surface roughness that influence surface wetting. To understand the role of surface chemistry and surface roughness in surface wetting, ALD was performed on several sets of cast polypropylene films, and results of contact angle analysis are shown in Figure 6b for sample sets prepared at 60 and 90 C. Deposition on the films (Figure 6b) and fibers (Figure 6a) was performed simultaneously. Contact angle for Al2O3 ALD on native-oxide coated silicon wafers is also shown. After ∼20 A/B cycles, the aluminum oxide ALD coating on SiO2 results in a contact angle between 10 and 20 when measured soon after deposition. It is noted that over several hours of air exposure the contact angle for aluminum oxide on silicon was found to increase, due most likely to an increase in surface carbon as observed in XPS analysis. The starting polypropylene film shows a contact angle of 120. It is important to note that even though these cast polypropylene films Langmuir 2010, 26(4), 2550–2558
are referred to as “planar”, they have inherent roughness that will affect the measured contact angle. The spread in the data, especially in the initial deposition regime, likely results from film nucleation and sample-to-sample variation in local roughness. The values obtained, therefore, are an approximation to the contact angles expected on a surface that is significantly more planar than the highly random textile surface. On the cast polypropylene, the advancing contact angle decreased slowly for the first 40-50 A/B cycles and then decreased more rapidly before saturating. The receding contact angles showed more irregularity but generally indicated larger hysteresis with increasing number of cycles. The initial slow decrease in advancing contact angle with film coating is ascribed to aluminum oxide nucleation and a mix of polymer and alumina nuclei present on the surface. It is interesting to note that after further deposition the contact angle on the cast film does not approach the value for ALD aluminum oxide on SiO2, even after 200 A/B cycles. This difference in contact angle for ALD coated polypropylene and SiO2 could be due to (1) incomplete coverage of the polypropylene by the ALD coating, even after 200 A/B cycles and/or (2) roughness of the polypropylene film surface. The results also suggest that the details of the surface evolution depend on substrate temperature. X-ray Photoelectron Spectroscopy. Given the observed effect of adventitious carbon on the contact angle of Al2O3 on silicon substrates, it is necessary to understand the effect of surface carbon species (from the polymer) on the measured contact angle. The relative intensity of the carbon 1s (C 1s) signal at ∼285 eV measured by X-ray photoelectron spectroscopy was used to characterize the fraction of carbon in the near-surface region of the fibers. At 1000 eV, the electron inelastic mean free path is ∼1-2 nm, so for the nonplanar samples studied here, the majority of the detected photoelectrons come from the top 1-3 nm of material.37 A silicon wafer with native oxide coated with 50 A/B cycles at 75 C showed a C 1s signal corresponding to ∼20 at. % after air exposure, and the C 1s signal decreased to
